Procedure:

You will have to take some data from some known secondary sources $($which need to be mentioned in the proposal$/$report$)$ and perform forecasting. The project proposal, ideally not more than $2$ pages long, should contain a brief description of the dataset you are going to use for the project, and some idea about what you are planning to forecast, the forecast horizon, and how you plan to develop the forecasting model.

The dataset is supposed to be divided into two parts $($in terms of no$.$ of data points$)$ in the ratio of $4$:$1($or $3$:$1$ or whatever sounds logical$).$ The first part, the training data, will be used for optimizing the model parameters and the second will be used only for the validation of the model.

You are free to use any algorithm $($or even multiple algorithms$)$ including those that haven't been covered in the class. The main content of the report should be around $15$ pages $($Times New Roman, $12$ font size, $1\mathrm{.}5"$ line spacing$),$ excluding plots, tables, and codes. The report should be code$-$agnostic: codes should appear as an appendix.

The evaluation will not be done only on the basis of accuracy in the forecast generated but on the logic behind using any particular method to forecast. It will be evaluated on the following aspects:

Criteria for Evaluation of the individual projects

Criteria Weightage

$$ Accuracy$/$Completeness of the work done $30\%$

$$ The complexity of the Problem studied $30\%$

$$ Presentation of the report $15\%$

$$ Logical Consistency of the steps used $15\%$

$$ Anything Extra i$.$e$.$ beyond what we have covered $10\%$